ABSTRACT Adults with traumatic brain injury (TBI) have chronic deficits in cognitive-communication and these impairments have been linked to negative outcomes and poor community reintegration and independence. A significant challenge in the field is that commonly used methods to detect these deficits in clinical and research settings lack the required sensitivity and have focused on a limited subset of discourse tasks that do not reliably predict communication outcomes. At the heart of our proposal is the idea that current theories of cognitive- communication are too narrow and the methods used to detect deficits are too limiting. In contrast to current conceptualizations which state that cognitive-communication deficits affect discourse and conversation (leaving basic sentence level processing intact), we propose that cognitive-communication impairment is a deficit in the flexible use and processing of language that manifests across the varied and dynamic contexts of everyday language use, whether processing a single sentence or participating in a multiparty conversation. For example, individuals with TBI make inappropriate or irrelevant comments, suggesting a failure to consider the communicative setting and the perspective of their conversational partners, as well as insensitivity to contextual cues (environmental, partner) that guide language use. These are problems using and processing language in context. From this perspective, there is a striking disconnect in the field between clinical observations of impairments in using language in context and the widespread use of decontextualized tasks and measures to capture these deficits in the lab and clinic (e.g., monologue discourse task). Using ecologically valid language tasks and methods sensitive enough to detect even subtle, though meaningful, disruptions in language, the proposed work endeavors to show that deficits in contextual language processing in TBI extend even to the rapid processing of individual phrases and sentences in the moment (i.e., online processing), can be captured in controlled settings, and that such deficits predict communicative outcome. The proposed program of research represents a novel direction in the study of TBI with substantial basic science and clinical translational significance. The proposal is organized around three AIMS: (1) To investigate language processing in context in TBI; (2) To investigate language use in group settings in TBI; (3) To determine the relationship between language processing in context and communicative outcome. This proposal is unique in the field and uniquely promising for understanding the nature of deficits in contextual language processing following TBI and, ultimately, improving rehabilitation intervention outcomes.